Sample-size requirements for developing strategies, based on the pupal/demographic survey, for the targeted control of dengue

Several methods to determine the sample size required for a reliable and practical assessment of the number of Aedes aegypti pupae in a community in Puerto Rico have been explored. Because the pupae were highly aggregated, the data were fitted to a negative binomial distribution. Classical statistical-inference methods for sample-size determination demanded the sampling of > 3000 premises for a reliable estimation of the mean number of pupae/person (with a 15% error). This number was reduced to 1000-1200 premises after applying a finite-population correction. Database sub-sampling simulations, with increasing sample sizes, showed that the variability in the mean relative abundance of container types and in the mean number of pupae/container substantially decreased after sampling 186 and 310 premises, respectively. Sequential sampling was applied to test the hypotheses that the number of female pupae/person was at least 0.19 (considered the dengue epidemic threshold) or no greater than 0.10 (arbitrarily set as the safe level). After sampling only 25 premises in the first survey and 125 in the second, it was determined that the densities of female pupae were above the epidemic threshold. Thus, sequential sampling provided substantial reductions in the sample size required to determine if vector control was needed. Validation of the Ae. aegypti thresholds required for dengue transmission could confer viability and efficiency to dengue-vector surveillance and control programmes.